Measurement for a web that stopped linking back.

Buyers used to search. Increasingly they ask, and an assistant answers with a shortlist. When someone acts on that recommendation, the visit usually arrives with nothing attached: no referrer, no query, no campaign tag. It lands in your reporting as “Direct”, which is not a channel at all — it is the bucket that means we don't know.

That gap is expensive in a specific way. AI-referred visitors are not average visitors; they arrive already recommended, and every independent dataset published so far finds them converting above non-branded organic search. The traffic you cannot see is likely your best traffic, which means the work that produced it cannot be justified, funded, or repeated.

Traceten closes that gap without reopening a worse one. We identify which assistant sent a visit — using signals other than the referrer, and reporting how confident we are rather than asserting a label — and we join that visit to revenue recorded in Stripe and Shopify.